一种基于多熵阈值图像分割的边缘检测新方法

该文提出的熵阈值分割图像边缘检测新方法，在噪声严重的情况下，能够提取出准确连续的物体边缘，实验结果表明这种方法能获得比Ｓｏｂｅｌ算子及Ｋａｐｕｒ多阈值分割法更为满意的边缘检测结果，这一方法在出版系统中具有一定的实用价值。

A New Method to Extract Edge Using Multiple Entropy Thresholding

In the field of image processing, edge extraction is an important subject which is the foundmention of some other further approach. The new method promoted here extract edge using multiple entropy thresholding and some special means to substract noises as well. So, it is able to detect accurate and continuous edges among serious noises. Experiments proved that this new method can obtain more satisfactory results than the Sobel's and the Kapur's thresholding method.